TSR2

by Damien Burke

  1

  XR219

  2

  XR220

  3

  XR221

  4

  XR222

  5

  XR223

  6

  XR224

  7

  XR225

  8

  XR226

  9

  XR227

  Stage 2 assembly jigs under construction at Weybridge in October 1961. These workers are using an optical sighting device to mark the site of the next jig in the sequence precisely. BAE Systems via Brooklands Museum

  Meanwhile, the Cold War was hotting up. An invasion of communist Cuba had failed in April after President Kennedy lost his nerve and called off supporting USAF and United States Navy air strikes while the US-trained Cuban-exile commandos were still at sea on their way to their target. This set the scene for closer Soviet–Cuban relations, and would eventually lead the world to the very brink of nuclear war. Tensions increased throughout 1961, and on 13 August that year the Soviets closed the border posts in Berlin, and within days began construction of the infamous Berlin Wall.

  On 1 September a nuclear explosion bloomed over the Semipalitinsk test site in Kazakhstan, marking the end of Soviet participation in an atmospheric test ban and the beginning of an intensive series of nuclear tests, climaxing with the detonation of the largest nuclear weapon ever produced, the ‘Tsar Bomba’, with a yield estimated at 57 megatons (the bomb having been dialled down from its maximum yield of 100 megatons). Had such a bomb been dropped on 10 Downing Street, all of London would have ceased to exist. The zone of total destruction would have extended beyond the modern-day M25 motorway and would have included Hatfield, Brentwood, Sevenoaks, Reigate and Woking. Bomber Command’s headquarters at RAF High Wycombe in Buckinghamshire would probably have been seriously damaged. The fireball and mushroom cloud would have been visible from anywhere in the British Isles. The Prime Minister, Harold MacMillan, suffered nightmares and wrote in his diary ‘… the last Russian tests are rather alarming’, and ‘… the one hundred megaton is not just a stunt. It would scorch with fire half France or England if dropped. What then should we do?’ It was not, it seemed, a matter of if the TSR2 would be needed to carry out is primary task, but of when.

  Multi-layered management

  From the outset the TSR2 project was burdened by multiple layers of bureaucracy, all of which helped to introduce further delays. The customer, the RAF, had its own bureaucracy in the form of the Air Ministry, and had written the requirement. Above that was the MoA (the MoS before 1960), which had ultimate say when it came to technical instructions and was also in charge of contracts. When it came to actual funding, however, authority had to be sought from the Treasury.

  On the TSR2 project, instead of simply giving the contract for the complete weapons system to a single company, it was decided to divide the project into different categories of equipment. In Category 1 were major systems for which the MoA would be directly responsible in terms of technical direction, the awarding of contracts, and the monitoring and control of costs. Vickers and English Electric would have no control whatsoever over these important items. They included the engines and most associated equipment, almost all of the nav/attack system (including terrain-following radar (TFR)), the reconnaissance pack, the complete communications and weapons carriage and release subsystems, ground servicing equipment, and the flight simulator. Category 2 contained the crumbs from the MoA’s table, for which Vickers would be responsible for technical direction and the award of contracts – albeit subject to Ministry approval of both chosen subcontractor and the contract terms. Category 2 items included the engine gearbox, ejection seats, AFCS, HUD, Doppler radar, SLR (but not its associated display unit) and the Central Computing System. A final category, Category 3, included everything else; basically anything that the Ministry did not consider important enough to exercise any control over, from nuts and bolts to complete airframe components. On this score at least Vickers had control right through from design to contract awards, though of course the Ministry reserved the right to keep an eye on any particular item and jump up and down if it considered Vickers was ‘not doing the right thing’.

  There was little apparent logic to the distribution of items within the categories, and each individual item was assigned a Director (one of seven, all reporting to an overall Project Director). To ensure that all of the disparate systems talked to each other once installed in the aircraft, an overall Systems Integration Panel was formed, consisting of each of the equipment directors plus representatives from Vickers, English Electric and the Air Staff. Below that were further Sub-panels, for example the Flight Control System Sub-panel. Vickers was held accountable at all times for the performance and cost of anything it subcontracted yet, when it came to anything under Ministry control, not only did Vickers have little say, but it also had no knowledge of the costs involved. It was a weighty and cumbersome management plan, but at least it was put together with the best of intentions.

  TSR2 procurement chain. Damien Burke

  A sealant test tank under construction. This incorporated the most difficult structural joints and was designed to show up any problems before work began on actual aircraft components. BAE Systems via Brooklands Museum

  Production

  The British Aircraft Corporation was justifiably proud of the modern methods employed, not only in manufacture of the airframe but also for the jigs. Optical sighting ensured accurate placement (absolutely vital with components being built not only on separate sites, but also at separate companies), and jig borers were controlled by a semi-automatic method based on punched card or tape readers. Along with the various jigs used in production, a variety of testing jigs were also erected, such as the large and ingenious dynamic fuel test rig built by Napier. Large enough to hold a complete fuselage with wing attached, this had provision to allow the entire assembly to be rotated in both the pitching and rolling planes. Thus the complete installed fuel system could be tested for leaks in various attitudes.

  Forming intake tunnel skins over a concrete former. Many of these formers survive to this day at the Brooklands Museum, close to the former Vickers-Armstrongs factory at Weybridge. BAE Systems via Brooklands Museum

  The Weybridge skin milling shop in October 1961, with various TSR2 integrally stiffened panels ready for final polishing. BAE Systems via Brooklands Museum

  Before starting production of components destined for the first airframes, materials tests were undertaken to assess the suitability of the chosen materials, and build tests were undertaken on example joints and sub-structures to assess the practicalities of putting everything together, and the overall strength of joints. Skin panels were mostly milled from solid billets of aluminium alloy, chemical etching being used on panels that had to be stretched or pressed to shape. Unlike modern-day computer-controlled milling, all such work on the TSR2 was carried out using manually controlled copy routing machines. Wood and rubber patterns were mounted high above or alongside the metal being worked on, and the operator guided the machine along the pattern part while the cutting head mirrored his efforts. Time-consuming polishing followed, initially down to 32 micro inches to remove all evidence of cutting marks, though economy measures were soon introduced to polish down to 60 micro inches, sufficient to remove cutting discontinuities though some cutting marks would remain visible. (These measures, however, did not get communicated to all production facilities, and such waste of effort continued throughout the project’s life.) The Skin Milling shop at Vickers Weybridge, the largest in Europe at the time, was simultaneously working on TSR2, VC10 and Vanguard skin sections, while English Electric’s workshops at Preston and Accrington produced fuel tank panels, engine-tunnel skins and tailplane spigots among other machined items. Shot-peening machines at Vickers Weybridge and English Electric Samlesbury enabled milled skins to be formed to match templates and, for example, turn flat milled planks into curved engine-tunnel skins.

 
The first airframe components began to be put together during November 1961, though shortages of particular materials (particularly titanium for bolts) were already delaying progress. The correct grade of steel for the undercarriage legs, which were being built under subcontract by Electro-Hydraulics Ltd, was also proving difficult to obtain, and consequently the first airframe would have its undercarriage built from a lower grade of steel. There were also delays in putting jigs into full production use, because workers had to be trained in their use and the training took longer than expected. Of almost as much concern as the internal delays was the fact that news of external delays at subcontractors was not reaching BAC for weeks or months after the problems had arisen. Because subcontracts were mostly placed by the MoA rather than directly by BAC, there was simply no channel for communications between BAC and most of the subcontractors. As a result, problems inevitably snowballed and gave BAC little or no opportunity to seek or recommend alternatives. This would be a recurring issue throughout the project, and meant that the prediction of a first-flight date in March 1963 was hopelessly optimistic.

  Machining one of the main fuselage frames from a solid billet of alloy. The guide frame above the work piece enabled the operator to produce identical components time after time, but it was still a time-consuming and skilled task. Note the lack of ear and eye protection. BAE Systems via Brooklands Museum

  The TSR2 construction sequence as drawn up in September 1960. This was carefully designed to cope with the work split between Vickers and English Electric, with the largest chunks of airframe being able to be transported by road without the need for special measures such as removing lamp posts en route. Stages 3 and 4 were eventually carried out in parallel, and the rear fairing became a task for BSEL rather than EE, though BAC was going to take it back in-house for production aircraft. Damien Burke

  Stage 1 forebody construction at Weybridge. The nose section was initially constructed in two halves, split along the aircraft’s centreline. BAE Systems via Brooklands Museum

  Stage 1 centrebody construction at Weybridge, August 1962. These are the intake tunnels, leading aft into the engine tunnels. The aft half of the centrebody was one of the most complex and highly stressed areas of the aircraft, containing the main wing attachment frame, undercarriage, bomb bay and fuel tanks. BAE Systems via Brooklands Museum

  Weight was also continuing to rise as more detailed design work ‘filled in the gaps’ with regard to exactly what was going in the airframe. By November 1961 BAC was predicting a weight of 94,258lb (42,780kg), and the MoA was predicting that this would mean an ‘in-service’ weight of 97,700lb (44,350kg). If the latter figure was accurate, the aircraft would not meet the requirement in two areas. The take-off run for the 1,000nm sortie would be 1,065yd (975m) rather than 1,000yd (915m), and the landing run would be 625yd (570m) instead of 600yd (550m). This, for the time being, was of no great concern to the RAF’s Operational Requirements team, as there was ‘nothing that [could] usefully be done about it’.

  Problems at BSEL were not limited to engine development (for which see Chapter 8). By mid-1961 English Electric, responsible for the rear fuselage, was becoming concerned about BSEL’s lack of progress with rear-fairing design and production. English Electric invited BSEL to send men up to Warton to complete the design and get things moving, but BSEL did not have the men to spare, and it soon became clear that it was out of its depth when dealing with the stressing and structural design requirements of a chunk of airframe. As engine problems mounted, the rear fairing suffered one delay after another. This was particularly galling to English Electric, which had demonstrated to BSEL the various manufacturing methods used at Preston, because BSEL had ignored them all, choosing more complex and expensive methods in just about every area in a bid to save weight on the finished item at all costs. The fairing continued to be a problem throughout the project’s life.

  By March 1962 production of the airframe had also slowed dramatically. So much, in fact, that progress on certain jigs was lagging up to five months behind the original schedule, though other components were a mere two months behind. Much of this was down to the steeperthan-expected learning curve for putting together this most complex of aircraft structures, but BAC was confident that airframes after the first one were going together more quickly, and that it would have regained much of the lost time by the time the fifth aircraft on the line was complete. However, the delays on the first aircraft were unlikely to get any better, and the first-flight prediction was adjusted, first to June 1963 and then to August 1963. As more and more drawings had been produced to fill in the details of various parts of the airframe, the estimated weight of the completed aircraft had also proved to be overoptimistic. By now it had risen to the region of 96,000lb (44,000kg), just past the point of being able to meet the performance specifications fully, and BAC was struggling not to increase it. Almost half of the weight gain was due to changes in the rear fuselage.

  The RAE, which had objected to some aspects of English Electric’s P.1 design some years earlier (resulting in the creation of the Short SB.5 research aircraft, which thoroughly vindicated English Electric’s choices), now set itself up to contribute another expensive mistake to the overall programme. The RAE had convinced Vickers of the need to build some scale models of the TSR2 airframe that could be launched by rocket to high speeds (up to Mach 1.8) to verify and expand on windtunnel results. Much of the TSR2’s windtunnel testing had been carried out using modified P.17 models, or using more representative TSR2 models after the design had been frozen, and these had revealed a possibility of some worrying stability characteristics. It was not clear, however, how much the constraints of the windtunnels and the supporting structures on which the models were mounted were affecting the results. The RAE’s plan was to verify these results by launching heavily instrumented 1/12th-scale models into the air. Two of the major areas of investigation were yaw stability at high Mach numbers and the effect of incidence at high speeds, using small pulse rockets fitted to the models to try to excite oscillations in these directions and measure the results. The secondary intention was to see what effect a sudden engine loss would have at high speed, to be simulated by a door partly closing off the intake on one side.

  Once these expensive models were available, Aberporth became the scene for a farcical series of tests that started in February 1962 and continued until February 1964. During these tests, seven of the eight model launches were either partly or entirely unsuccessful owing to malfunctioning rockets or instrumentation, or to the models settling into spins (their dynamic characteristics did not match those of the real aircraft, particularly in roll, and the tailplanes were set to make each model to fly a gentle barrel roll in order to try and keep the model within the test range area). After launch the rocket would fire for three seconds and then fall away, allowing the model to fly free and decelerate while measurements of pressure and acceleration were recorded, along with Doppler signals from the model. The models were also tracked on ground-based radar. In the end only one test provided really useful results, recording while the model decelerated from Mach 1.48 to Mach 0.78. Every test above Mach 1.5 produced a spin, induced by the model’s mismatch to the aircraft’s dynamics. Overall, the programme was considered by English Electric’s windtunnel team to be little more than the production of some expensive fireworks, and not a good advertisement for the RAE, which hitherto had been redeeming its somewhat tarnished reputation at Warton with some sound theoretical work. Shortly after the tests had begun, further windtunnel work by the Aircraft Research Association at Bedford had pinpointed most of the problem areas anyway, though finding and implementing practical aerodynamic solutions at this late stage was not going to be possible (the AFCS’s manoeuvre boost system would end up being modified to compensate).

  Royal Aircraft Establishment personnel at a chilly Aberporth load the first TSR2 free-flight model on to a rocket in February 1962. This particular model’s configuration was quite unlike that of the final TS
R2, exhibiting a small fin and sawtooth leading edge on the wing, and was therefore useful only to prove the launching technique and instrumentation. Rocket and instrumentation malfunctions blighted the free-flight test programme. BAE Systems via Warton Heritage Group

  One of the RAE’s free-flight TSR2 models is launched by rocket booster. Many such launches ended up with the model settling into auto-rotation and providing no useful data. More successful results were achieved with free-fall spinning models that were simply dropped. BAE Systems via Warton Heritage Group

  Meanwhile, progress on the various third-party systems to be fitted to the airframe was as varied as that on individual jigs at Vickers/English Electric. For example, Elliott’s AFCS was coming along nicely, with a basic model ready to be incorporated into the integrated testing rig at Weybridge, but the power control actuators from Hob-son, a vital part of the AFCS, still showed no sign of arriving a full six months after Hob-son had predicted a five-month delay owing to supply problems with high-duty alloys. It did not help that, after initial wrist-slaps about timely communication of delays, subcontractors had now got into the habit of holding on to their own bad news until the last possible minute, in case another system elsewhere within the aircraft fell further behind schedule and thus let them off the hook, if only partly. The lack of firm control over the project as a whole was by now seriously jeopardizing any chance of meeting even the adjusted schedule, never mind the original one.

  Further needless delays were introduced by the MoA’s love of protracted contractual negotiations. Subcontractors, and even the main contractor, BAC, would often find themselves pressured to complete design studies into particular items of equipment, with a contract arriving weeks or months later to cover this work, but nothing being put in place to cover actual build work after that point. Instead, the design study would remain in limbo, often for months, while the MoA and the contractors argued over the finer points of the sort of contracts that had already been dealt with time and time again. Understandably, companies were reluctant to begin work without written assurance that they would be paid for it (a formal letter of Intention To Proceed at the very least), but, in many cases, had they not done so the overall project would have fallen so far behind as to guarantee its cancellation. This meant that companies sometimes took huge risks to keep the project on track, if they could afford to do so. Those that could not soon began to wish they had chosen easier customers, and the delivery date of their equipment slipped ever further away through no fault of their own.